T HE RUPESTRIAN CHURCH OF SAN PIETRO DA MORRONE

(MATERA, ITALY): INSIGHTS AND PHOTOGRAMMETRIC-BASED

CHARACTERISATION

Rafael Ramírez Eudave * 1

Institute for Sustainability and Innovation in Structural Engineering (ISISE)

University of Minho, Portugal.

Michele D’Amato, Giuseppe Andrisani

Department of European and Mediterranean Cultures (DiCEM)
University of Basilicata, Italy.

Keywords: rupestrian architecture, photogrammetry, geometrical characterisation,

three-dimensional modelling

DOI: 10.48255/1973-9494.JCSCH.21.2021.12

1. Introduction

Architectonic manifestations are vehicles for the creativity and the spirit of a cul-
ture. These assets represent documental evidence of human technical capacities, be-
liefs, identities and the vision that people have about themselves and the world. The
city, then, represents a process of permanent construction of heritage as one of the
cores for collective cohesion [1]. It is important to recognise the symbiosis between the
intangible and tangible values that coexist in a cultural asset. In other words, to identi-
fy and characterise the physical consequences of the cultural assets (such as colours,
geometries, spatial organisations, elements of design, etc.) is a necessary step for
contextualising the object and its intangible values.
Protecting the cultural heritage is an essential activity for ensuring the continuity of
a culture and its values; the presence of cultural assets is not sufficient by itself. It is
necessary to adequately transmit and share their meanings and values. In other words,
the social function of heritage depends on the socialisation and adequate interpreta-
tion of the material objects. The international recommendations on Historic Urban
Landscapes (HULs) [2] remark the value of documenting the historical city as a step
for promoting the knowledge of cultural assets. This process is important for sharing,
but also for the development of scientifically supported interventions, also in the con-
text of natural risks [3].
There are numerous examples of the use of photogrammetry for surveying cultur-
al assets in the continuous process of its enhancement [4]. Some examples include
the use of Unmanned Aerial Vehicles for facilitating data acquisition on territorial-scale
surveys, such as archaeological sites [5]; other relevant experiences integrate the gen-
eration of three-dimensional models of architectonic assets in the context of territori-
al-scale maps [6] and even the use of unconventional sensors, such as spherical pho-
tographs [7]. Moreover, the application of the photogrammetric approach for religious

* Corresponding author: r.92@outlook.es

 cultural assets has successful examples. The case of the Church of San Francisco in
Betanzos (La Coruña, Spain) demonstrates the three-dimensional modelling of a com-
plex gothic structure [8]. A remarkable example of photogrammetric survey, including
wall-paintings, can be found in the work on the churches in Göreme Valley (Cappa-
docia, Tukey) [9] or in the works of Russo et al. [10] in which this approach was used
for mapping the decay of historical façades.
In the context of a globalising world, there is an increasing interest in recognising and
protecting those cultural expressions that transcend their local contexts and become rep-
resentative for the whole of humankind. At an international level, the United Nations
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

Educational, Scientific and Cultural Organisation (UNESCO) encourages the recognition

of outstanding examples of cultural heritage, such as the case of the Sassi di Matera.

2. The «Sassi di Matera»

The city of Matera is located in the region of Basilicata, in the south of Italy (Figure 1),
a region that has been permanently inhabited since the Palaeolithic period due to the
convenient availability of several resources and a favourable geological environment
[11]. There are two well-defined limits framing the territory of Matera. The east side is lim-
ited by a canyon (Gravina della Murgia) that divides the city from the plateau of the Murg-
ia. The west side is limited by the channel of the Bradano river. This region represents
geomorphological particularities associated to its geological past.
The area emerged from the water in the Pleistocene age [12], exposing platforms of
sedimentary soils with a relatively high proportion of organic components. These depos-
its of calcarenite are the source of the rich variety of landforms in the region. In general
terms, the upper layers of the former seabed are relatively soft, favouring significant
erosion even from small currents of water, which has given rise to the formation of can-
yons, such as in the case of the Gravina della Murgia. This process thus exposed wide
surfaces of stone that were relatively easy to work, facilitating the anthropic configura-
tion of the territory by using the excavated spaces that had been formed (Figure 2).
This relatively soft stone, calcarenite – locally known as «tufo» (tuff), not to be con-
fused with the homonymous material of volcanic origin in the region of Campania – is
critical for explaining the development of the human settlements in the region. It is a
material with elevated porosity (up to 50%) and relatively poor mechanical properties
(from 1 up to 5 MPa of compressive strength) [13-15]. The composition of this materi-
al includes organic content, namely marine shells. The availability and workability of
this material permitted the building tradition in the region to continue for a long period,
resulting in a permanent improvement of the architectonic typologies. Most construc-
tions were originally based on the opening of excavated habitable spaces by extract-
ing tuff from the walls of the Gravina della Murgia. Some constructions were later com-
plemented by using the extracted material for exterior walls, façades and even
enlargements.
These architectonic dynamics configurated a very particular landscape, determin-
ing an urban organisation that also reflected the consolidation of social behaviours. For
example, a remarkable characteristic of Matera is the existence of small clusters of
houses (“vicinato”) that shared a semi-public space which was used for collective ac-
tivities, namely related to water supply, social contact and as a release from the con-
straining excavated spaces. Another relevant attribute is the existence of collective wa-
ter reservoirs for storing rainwater. The historical city of Matera (also known as the

260
Figure 1. Location of the Region of Basilicata in Italy. (a) Location of Matera in the Region of
Basilicata and; (b) satellite view of the city and; (c) Gravina della Murgia. (source images: (a) and
(b) Open Streen Maps®; (c) Google Maps®).

CONSERVATION SCIENCE IN CULTURAL HERITAGE

Figure 2. Sassi di Matera as seen from the west.

Sassi di Matera) materially represents a long tradition of knowledge on sustainable so-

cial dynamics [16].
The first decades of the 20th century in Matera were characterised by sustained
demographic growth that coincided with an important pastoralist crisis. The situation
caused a marked degradation in living conditions that resulted in the violent interrup-
tion of the continuity of Matera’s urban life after the Second World War [17]. The com-
bination of events resulted in a substantial overcrowding of the rupestrian housing
units, leading to poor sanitation and hygiene; the lack of sewage and running water se-

261
 riously compromised the delicate equilibrium between the city and its resources [18].
The precarious living conditions, largely denounced by authors such as Carlo Levi,
drastically impacted public policies. In May of 1952, the inhabitants of the Sassi were
displaced to new neighbourhoods in the surroundings of the historic city. The Sassi,
consequently, became a ghost town that suffered a continuous process of degradation.
It was not until 1986 that citizens were able to return to the historical city, amid
growing interest in the city and its cultural values. Those values were internationally
recognised when UNESCO declared the Sassi a World Heritage Site in 1993 [19].
More specifically, the ensemble of the Sassi di Matera was recognised as an outstand-
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

ing example of continuous human appropriation of the landscape [20]. The declaration
states that Matera meets three of the World Heritage criteria [21]:
“Criterion (iii): The Sassi and the Park of the Rupestrian Churches of Matera repre-
sent an outstanding example of a rock-cut settlement, adapted perfectly to its
geomorphological setting and ecosystem and exhibiting continuity over more
than two millennia.
Criterion (iv): The town and park constitute an outstanding example of an architec-
tural ensemble and landscape illustrating a number of significant stages in hu-
man history.
Criterion (v): The town and park represent an outstanding example of a traditional
human settlement and land-use showing the evolution of a culture which has
maintained a harmonious relationship with its natural environment over time”.
One of the consequences of the gradual return of urban life to the centre of Matera
is the rediscovery of places that were subsequently abandoned and, in some cases,
forgotten. The inclusion, study and rehabilitation of these constructions is a necessary
step toward their successful urban regeneration. In fact, the vision of a sustainable his-
toric city must start with the correct contextualisation of its assets, recognising in their
configuration a testimony of social dynamics, and the use of available resources and
technical capacities [22]. A remarkable case of a rediscovered cultural asset is the an-
cient church of San Pietro da Morrone.

3. The church of San Pietro da Morrone

The object of the case study is located in the district of Civita, the sector that corre-
sponds to the former fortified city and is found at the intersection of two streets - Via Ci-
vita and Via San Potito. The church of San Pietro da Morrone is an example of an en-
tirely excavated construction, a frequent typology in the old city of Matera. This building
was explored in 2013 after more than four decades of abandonment [23]. The research
work by Fontana and Paolicelli [24] after rediscovering the building gives an insight into
more than seven centuries of history associated with the building.
The first mention of the church is in the testament of the Constable, Angelo de Be-
rardis, dated 30th May 1318. It is mentioned as one of the main churches in the city. It
is later mentioned as the Temple of San Pietro De Cataldis in the context of the visit of
the Cardinal Michele Saraceno (1543-1544) and, as San Pietro de Morronibus in the
context of the visit of Mons. Vincenzo Giustianini (1595-1596), being depicted as aban-
doned and degraded at that time. In 1623-24 the temple is described as the property
of the Zaffari (or Cataldi) family. The existence of damage resulting from water infiltra-
tion and the need for urgent repair work is also documented.

262
 The building is not mentioned in the list of churches of the archbishop of Acerenza
and Matera, Vincenzo Lanfranchi (1667), which may indicate that the building was not
being used as a place of worship at that moment. In 1701, Domenico Antonio de la
Torre, archbishop of Matera [25], describes the church as profaned, deconsecrated
and full of debris. He tries to buy the construction, describing it as a building in two sec-
tions. The temple is later mentioned as a cellar (1740), under the name of San Petuddo.
The documents of the notary, Belisario Torricella (1774), describe a centuries-old church
with the corrupted name of San Pituddo, reduced to being used as a cellar and devoted
to a saint named San Pedrino in Tuscan dialect.
The cadastre of 1816 includes the property as a housing unit. By 1868 it is men-
tioned as the cave and church of San Potito and Petullo. The church was sold (under the
concept of a house) in 1869 and was sold once again two years later. The Urban Ca-
dastre of Matera of 1878 associates the property to the parcel number 1203 in Via Ci-
vita 1. The property was inherited in 1905 and 1937. The last testimonies are the acts
of 1960, that declare the construction to be uninhabitable. The church, at that time as-
sociated to the numbers 19 and 20 of Via San Potito, was finally closed on the 5th June
1961 under Law 619 (1952). The corresponding act includes a schematic sketch of the
layout (Figure 3).

CONSERVATION SCIENCE IN CULTURAL HERITAGE

Figure 3. Schematic sketches of the floor plan produced for the 1960 acts (source: adapted from [24]).

The building has an irregular floor plan, which is typical in constructions in Matera.
In fact, the workability of the rock usually constituted a basis for guiding the growth of
a construction. The area of the temple is about 116 m2, distributed along two main gal-
leries. The construction is partially located under Palazzo de Vitis, a construction from
the 17th century (Figure 4). The façade of the church presents two accesses, with no
distinctive elements. There are only two small windows that may correspond to a for-
mer kitchen. At the moment of its rediscovery, there was a large amount of debris in the
church, obstructing some sections of the building.
As most of the rupestrian buildings in Matera, the geometry of the church is not only
irregular in the distribution of spaces: the vaults and walls are geometrically complex.
Furthermore, there are several structural reinforcements and additions to the original
construction, such as masonry arches, that seem to have been added to the original
vaults. Interventions of reinforcement may have been carried out because of the add-

263
 ed weight when Palazzo de Vitis was built (Figure 5). The presence of these reinforce-
ments and other modifications demonstrates the town’s dynamics in adapting and con-
tinuously reconfiguring the urban environment.
The geometric characterisation of the building was a challenging task. One of the
approaches that seemed suitable to accurately characterise the church was the gen-
eration of digital three-dimensional models based on point clouds. This approach has
been found especially useful when used for digitalising archaeological entities, due to
the convenient cost/benefit ratio. It also facilitates the obtention of accurate measure-
ments avoiding contact with sensitive surfaces.
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

Figure 4. Oblique view showing the façade of the church (1), the access from Via San Potito (2),
Via Civita (3) and the Palazzo de Vitis (4) (source: Image (Data SIO, NOAA, U. S. Navy, NGA,
GEBCO) Google Earth ®).

Figure 5. Image of the façade of the church (a) and its surrounding context, including Palazzo
de Vitis (b).

264
 4. The photogrammetric approach for San Pietro da Morrone: materials and
methods

The use of the laser scanner is a very precise technique and has become a regu-
lar method for surveying cultural assets. However, it demands a relatively high invest-
ment in equipment and relatively complex post-processing work. On the other hand,
the photogrammetric approach enables reasonably reliable results to be obtained by
means of relatively simple equipment for acquiring data combined with relatively us-
er-friendly post-processing computing. The laser approach is based on an active de-
vice that creates a cloud in which each point is associated to a precise measurement,
while the photogrammetric processes are based on passive sensors (namely a photo-
graphic camera), to obtain overlapping images which permit the identification of the
relative changes in some regions of the pictures in different frames.
The use of analogical photography for reconstructing three-dimensional objects has
been widely used in the 20th century, in many cases to obtain information from aerial pho-
tographs taken from planes for use in mapping, distinguishing land features, and so on.
However, it is possible to use very similar principles for reconstructing objects of a more
reduced scale, such as buildings. This technique, called “close-range photogrammetry”,
facilitates the obtention of sets of relative measurements that can be scaled for process-
ing real-scale three-dimensional models [26]. Since the relative changes in different
planes correspond in proportion to the distance between those planes and the camera,
it is possible to calculate a relative coordinate for these distances in multiple planes, cre-
ating a three-dimensional cloud of points that maintains the proportions of the object [27].
While the laser scanner technique provides a point cloud that represents the true size
and volumetry of the object, the photogrammetric approach provides a volumetry only
with true proportions but is able to be scaled in order to have the real size of the object.
The calibration of the cameras, however, permits to accurately scale the model based on
references, depending on the specific needs of the modelling process [28].

CONSERVATION SCIENCE IN CULTURAL HERITAGE

The quality of the outcome of the photogrammetric approach depends on the data
acquisition procedures as well as an adequate post-processing of the images. The ba-
sic procedure consists of taking as many photographs as needed for covering the
whole surface of the object. The images must satisfy a set of conditions, such as hav-
ing relevant overlapping (more than 60% is needed for close range photogrammetry),
not having any substantial changes in general light conditions and maintaining the
same focal length.
The redundancy on the photographs (i.e., the overlapping between contiguous im-
ages that guarantee that a certain surface appears in more than one photograph) is
critical for further point matching. It is important to consider the geometry of the sur-
veyed object in order to decide the most suitable strategy for scanning the surface
during data acquisition [29]. The present case of study favoured the acquisition of im-
ages as a panorama from a central point. Depending on the scale of the survey, it is
possible, and desirable, to include a series of complementary equipment for field work,
such as light sources, reflectors, tripods, supports, etc. When using digital cameras for
the data acquisition process, it is desirable to store the images in a minimally pro-
cessed format, such as a RAW file format. Even if these files are significatively larger
than the regular formats, they give better results during the matching process.
There are numerous commercial, open-source and free software options for gen-
erating point clouds from photographs. Besides, it is possible to manipulate and edit
the generated point clouds by means of Computer-Assisted Design (CAD) software

265
 and numerous digital modelling platforms. It is important to consider which tools will be
used to design a suitable and consistent workflow.

4.1. Data acquisition

There were several limitations in acquiring data for the church of San Pietro da
Morrone, namely due to the difficulties for scheduling visits to the building. Due to this
situation, only one session for data acquisition was performed. The passive sensor
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

device was a Canon® EOS Rebel T3 camera equipped with an 18-55mm lens. A tripod
was used for more stable images and a 30-watt LED lamp was used during the data
acquisition. However, this light source was limited to the size of the building, making
it necessary to use a fixed ISO of 1600, and speeds around 1/60 s with a lens aper-
ture of f/5.6. All the pictures were primarily stored in RAW file format with 4272 x 2848
pixels resolution and a pixel size of 5.34 x 5.34 um, giving individual files of approxi-
mately 20 MB each.
The methodology for photographing the building started with a plan of the route to
scan the walls and ceilings of the construction completely. A total of fourteen stations
– corresponding to a total of twelve inner spaces and two exterior stations (Figure 6) – were
proposed, based on a first on-site visual inspection and the sketches of the 1960 acts.

Figure 6. Planned route and stations. Each station corresponds to a spatial compartment of the
construction.

Each station was used as a fixed point for taking vertical sweeps of the walls and
ceilings, rotating the camera just a few degrees between strips. Each photo was com-
pared with the previous one in order to guarantee no less than 60% of overlapping. A
total of 1001 pictures were necessary to cover the inner surface of the church. Howev-
er, only 972 individual pictures were considered as adequate. The excluded pictures
were mostly overexposed, unfocused or presented shadows or undesirable marks due
to the presence of dust particles (Figure 7).

266
Figure 7. Discarded photographs due to over-exposure of regions (a) and (b), or unfocused
regions (c).

4.2. Data processing

In order to facilitate the processing of the pictures, they were divided into three sets
(chunks) corresponding to the main divisions of the church and the stations: stations 1,
2, 3, 11 and 12 (646 images, set 1); stations 4, 5, 6 and 7 (220 images, set 2); and sta-
tions 8, 9, 10 (106 images, set 3). The location of each station is shown in Figure 6.
Each set was individually processed by using a 30-day trial version of the Agisoft Pho-
toscan software, version 1.2.5.2594. This software is easy-to-use and intuitive, making
it simple, even for non-experts, to perform semi-automatic processes.
The software offers a standardised workflow, suggesting a series of default parame-
ters. The first process consisted in the alignment of the photographs. The pair preselec-

CONSERVATION SCIENCE IN CULTURAL HERITAGE

tion was disabled. This pre-process tool is intended to identify overlapping areas based
on downscaled versions of the images but may induce errors when analysing complex
geometries or surfaces with no distinctive elements. A high accuracy was selected, so
even some difficult overlaps were recognised through a more intense (but slower) itera-
tive process. The default key point limit (40,000) and tie point limit (4,000) were selected.
The outcome of this first process was a low-density point cloud giving an initial ori-
entation that was subsequently used for facilitating the creation of a dense point cloud.
The initial point cloud consisted of 1,006,132 points for Set 1; 626,622 points for Set 2
and 306,942 points for Set 3, giving a total of 1,939,696 points. A medium quality was
selected, with a moderate depth filtering, allowing the obtention of a dense point cloud
of 131,568,523 points.
The generation of a mesh was based on the dense point cloud and was made by
selecting an arbitrary surface type in the software, as well as a medium face count. The
default custom face count of 200,000 was selected and interpolation was enabled to
perform this process.
Texture was added to the mesh by selecting the generic mapping mode, using in-
formation from all the cameras and blending based on a mosaic model, with no colour
correction. The default texture size of 4,096 was selected. This last step provides a vi-
sual three-dimensional reconstruction of the surveyed space, with rectified images in
multiple planes (Figure 8). The final model was composed of a total of 6,686,337 faces
and 3,344,441 vertices.

267
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

Figure 8. Processing steps. Simple point cloud (a); dense point cloud (b); mesh (c) and surface
with photographic texture (d).

Figure 9. Volumetric mesh of the entire church. View of a section parallel to the ground plan (a)
and entire volume (b).

268
 The integration of the three chunks started with a manual alignment, orienting each
chunk on its relative position to facilitate the automatic co-registration of the sections
and the assembling of the entire model (Figure 9). Additionally, the three-dimensional
axes were rotated to have a coherent spatial framework.
The final result helps gain a better understanding of the geometry of the spaces
and the rock masses that divide the main galleries. This model also makes it easier to
observe the difference in levels and some subtle changes on the plane of the walls that
are not easily perceived on-site, namely because of the variable slope of the floor.
This comprehensive model was a valuable source for obtaining diverse profiles of
the construction, which permitted us to extract multiple plan views and elevations,
making it possible to measure and geometrically describe the numerous particularities
of the space.

4.3. Results and discussion

Several challenges emerged in managing the comprehensive model, namely due

to the limited capacity of the computer. For that reason, the original mesh of more than
one million longitudinal elements was reduced to a fifth of its original density. Neverthe-
less, no observable changes were identified, mainly because most of the eliminated
vertices corresponded to almost coplanar surfaces.
Once the mesh has been obtained, it is possible to externally manage the model in
numerous CAD platforms, such as Autodesk AutoCAD. Through simple volumetric op-
erations it was possible to section the entire volumetry to obtain rectified photos for
certain planes of interest. The orthophotos also permitted us to easily obtain a set of
drawings to support the development of an architectonic project (Figure 10).
Even if this church does not present large frescoes or ornaments on the walls, the
generation of orthophotos of longitudinal sections demonstrate the potential utility that

CONSERVATION SCIENCE IN CULTURAL HERITAGE

this technique would have in the context of decorated interiors. Besides, these images
are useful for documenting and analysing the pathologies found in the stones. Since the
orthophotos represent the real proportions of the building, they are useful for testing
project decisions and generating the corresponding documents (Figure 11). The use of
an ISO of 1600 for the images intrinsically added the presence of grain as a noise
source, which created minor artificial irregularities that are mostly visible in some re-
gions of the orthophotos. The model was scaled using the Scale Bar setting and the
largest measurements obtained in situ as a base. This procedure allowed verification of
the right correspondence between the proportions of the model and the construction.
The automatic report provided by the software gives practical insights on the pro-
cessing of the original photographs. The Ground Sampling Distance (given as ground
resolution by the software) was 0.704 mm/px, covering a total area of 116 m2. The to-
tal number of projections was 7,591,694 with a reprojection error of 0.725 px. The mean
key point size was 5.25 px with an effective overlap of 3.30 px.
The volumetric model makes it easy to clearly distinguish some stages in the con-
struction of the church, namely through the identification of geometrical features that
vary from one space to another along the two main axes of the church. A very interest-
ing visible element on this photogrammetric-generated plan is the existence of a very
thin wall between spaces 3 and 11, which would support the hypothesis of a pre-exis-
tent communication between these spaces, thus helping to reconstruct possible con-
figurations that the building potentially had in the past.

269
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

Figure 10. Photogrammetric reconstruction of the ceiling (a) and floor (b) plans.

From a structural point of view, the model was found useful for primarily identifying
reinforcement elements that probably correspond to strengthening actions carried out
during or after the construction of Palazzo Vitti, by projecting the perimeter of the Palaz-
zo onto the plan of the church. In fact, it is important to note that there are several arches
that would have specifically been built to bear the loads of the Palazzo (Figure 12).
The model is a valuable source for outlining an intervention project on this con-
struction. It is possible to obtain as many sections and plans as needed, facilitating the
integration of three-dimensional design processes as well. Since the surfaces are reg-
istered with the true photographic colours (as a mosaic), it is possible to perform accu-
rate mappings of biological decay (Figure 13). It is convenient to note, however, that
the photographic campaign on the building had to deal with some limitations due to the
presence of debris and temporary scaffolding. These objects appear in the model and
create some distortion in the real appearance of the building.

270
 CONSERVATION SCIENCE IN CULTURAL HERITAGE
Figure 11 Longitudinal section obtained by means of rectified photographs as a base for the
design process.

Figure 12. Section views contextualising spatial relations between Palazzo Vitti and the Church of
San Pietro da Morrone.

271
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

Figure 13. Projection of the Palazzo Vitti superimposed over the plan of the church (left); identifi-
cation of areas showing pathologies on the façade (right).

Figure 14. Selection of measurements obtained from the three-dimensional model.

Other relevant constraints were the difficulties associated with the illumination. It was
difficult to generate adequate light conditions using a single lamp. This difficulty was re-
flected in regions of shadow that were not adequately modelled. These regions are spe-
cifically located in some corners and represent holes in the three-dimensional mesh.
Even if the scale of these voids is almost neglectable in the context of the model, it is nec-
essary to recognise them as a limitation for the representativeness of the model.

272
 A set of measures obtained from the model is shown in Figure 14. When compar-
ing measures obtained on site against the ones provided by the model, no significa-
tive variations were identified, reaffirming the suitability of the model for describing
the construction.
It is convenient to note that even relatively small errors may accumulate, being es-
pecially difficult to identify on complex geometries like that of the church of San Pietro
da Morrone. These errors may become relevant distortions on the geometry when the
object is unproportionally long, for example. Even if it is acceptable to have a certain
degree of error, it is convenient to check if error is uniformly distributed throughout the
model or if there is a region or direction in which it becomes more relevant. A feasible
way for verifying this randomness and distribution of errors is by comparing different
distances on both the model and the building. A set of measurements performed for
this purpose is summarised in Table 1. It is remarkable that the larger relative differenc-
es are found on relatively short distances. This can be explained because of the very
irregular surface of the walls. The maximum absolute difference was of the order of
18cm in the longest axis of the church, of 16.53m (measured with a laser device). The
differences were considered acceptable as a first-approach representation, but the
model may not be suitable for designing detailed intervention projects.

Table 1. Measurements obtained in-situ and from the model.

Measurement Measurement with Absolute Difference

Measurement
from model (m) laser device (m) difference (m) (% of laser meas.)

(a) 0.51 0.55 0.04 7.27

(b) 0.45 0.44 0.01 2.27
(c) 3.5 3.45 0.05 1.44

CONSERVATION SCIENCE IN CULTURAL HERITAGE

(d) 3.65 3.65 0.00 0.00
(e) 2.09 2.09 0.00 0.00
(f) 1.31 1.31 0.00 0.00
(g) 1.98 2.00 0.02 1.00
(h) 1.89 1.93 0.04 2.07
(i) 0.92 0.88 0.04 4.54
(j) 1.20 1.19 0.01 0.84
(k) 1.02 0.99 0.03 3.03
(l) 0.46 0.50 0.04 8.00
(m) 2.19 2.19 0.00 0.00
(x) 14.68 14.52 0.16 1.10
(y) 16.53 16.35 0.18 1.10

5. Conclusions

The process of documenting and characterising cultural assets is not only mean-
ingful for their conservation and protection, but also for contextualising their impact in
wider entities, such as the Historical Urban Landscape. The Sassi di Matera are a re-
markable example of human adaptation to a particular geological environment. Fur-

273
 thermore, they represent a series of strategies for optimising the use of natural re-
sources for satisfying basic needs, such as housing. The use of excavated spaces for
solving a wide variety of needs resulted in unique cultural products, such as in the case
of the church of San Pietro da Morrone. The abandonment of the building and the en-
tire city centre in the middle of the 20th century interrupted a continuum of transforma-
tion and adaptation.
The rediscovery of the church represents an opportunity for testing contemporary
approaches for documenting and characterising cultural assets. A valuable technique
for geometrically characterising the built environment is to generate three-dimensional
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

models based on photographic information. These images are computationally pro-

cessed in order to identify common points and reconstruct the relative position of the
points in the space.
The experimental campaign accomplished the purpose of generating a three-di-
mensional model of the church. Several constraints were found during the experience,
namely due to the limited time for conducting the in-situ survey and the lack of ade-
quate light sources. Nevertheless, a set of more than 1,000 photographs was obtained
and it was used as a base for developing a three-dimensional model.
The proportions of the model were scaled in order to obtain a reasonably accu-
rate and descriptive representation of the volumetry of the church. This scaled mod-
el was the basis for numerous documents, such as plans, elevations and sections
which gave an insight into some of the characteristics of the building and, further-
more, its potential structural relations to Palazzo de Vitis. The level of accuracy, how-
ever, may be enhanced in order to generate a more refined model for supporting ex-
ecutive structural projects.
This experimental program has permitted the development of a valuable model that
can be further refined and enhanced. In fact, a more complete and detailed photo-
graphic campaign (including the use of on-site references and scales) would probably
give more reliable results and a better idea of the real dimensions of the construction.
This first experience, however, confirms the suitability of using the photogrammetric
approach for documenting and increasing knowledge of the rupestrian cultural assets,
thereby facilitating their conservation.

Acknowledgments

This research was funded by the Erasmus ELARCH (Euro Latin-America Partner-
ship in Natural Risk mitigation and protection of the Cultural Heritage) 552129-EM-
1-2014-1-IT-ERA MUNDUS-EMA21 Program.

References

[1] UNESCO (2014) Indicadores UNESCO de cultura para el desarrollo. Manual me-
todológico. Patrimonio, Relevancia de la dimensión para la cultura y el desarrollo.
Available at: https://es.unesco.org/creativity/sites/creativity/files/iucd_manual_
metodologico_1.pdf [Accessed: 12th July, 2021].
[2] UNESCO (2016) The HUL Guidebook, The HUL guide, p. 59. Available at: http://
historicurbanlandscape.com/themes/196/userfiles/download/2016/6/7/
wirey5prpznidqx.pdf. [Accessed: 12th July, 2021].

274
[3] Hadjimitsis, D.; Agapiou, A.; Alexakis, D.; Sarris, A. (2013) Exploring natural
and anthropogenic risk for cultural heritage in Cyprus using remote sensing
and GIS, International Journal of Digital Earth, 6 (2), pp. 115-142. doi: 10.1080/
17538947.2011.602119.
[4] Gonizzi Barsanti, S., Remondino, F. and Visintini, D. (2013) 3D surveying and
modeling of archaeological sites – some critical issues. ISPRS Annals of the
Photogrammetry, Remote Sensing and Spatial Information Sciences, II-5/W1(i),
pp. 145–150. doi: 10.5194/isprsannals-II-5-W1-145-2013.
[5] Senkal, E., Kaplan, G. and Avdan, U. (2021) Accuracy Assessment of Digital
Surface Models from Unmanned Aerial Vehicles. Imagery on Archaeological
Sites’, International Journal of Engineering and Geosciences, 6 (2), pp. 81-89.
doi: 10.26833/ijeg.696001.
[6] Doğan, Y. and Yakar, M. (2018) GIS and Three-Dimensional Modeling for Cultur-
al Heritages, International Journal of Engineering and Geosciences, pp. 50-55.
doi: 10.26833/ijeg.378257.
[7] Fangi, G. et al. (2018) Improving spherical photogrammetry using 360° OM-
NI-Cameras: Use cases and new applications, International Archives of the
Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS
Archives, 42(2), pp. 331-337. doi: 10.5194/isprs-archives-XLII-2-331-2018.
[8] Ramos, A. P. and Prieto, G. R. (2015) 3D virtualization by close range photo-
grammetry indoor gothic church apses. the case study of church of San Francis-
co in Betanzos (La Coruña, Spain) International Archives of the Photogramme-
try, Remote Sensing and Spatial Information Sciences - ISPRS Archives, 40(5W4),
pp. 201-206. doi: 10.5194/isprsarchives-XL-5-W4-201-2015.
[9] Higuchi, R., Sugawara, H. and Gülyaz, M. E. (2019) Photogrammetry using vir-
tual restoration of wall-paintings of the rock-hewn churches in the Göreme Val-
ley, Cappadocia and its value for the museum’s contents, ISPRS Annals of the
Photogrammetry, Remote Sensing and Spatial Information Sciences, 4(2/W6),

CONSERVATION SCIENCE IN CULTURAL HERITAGE

pp. 83–89. doi: 10.5194/isprs-annals-IV-2-W6-83-2019.
[10] Russo, M. et al. (2019) Modeling and deterioration mapping of façades in histor-
ical urban context by close-range ultra-lightweight UAVs photogrammetry Inter-
national Journal of Architectural Heritage. Taylor & Francis, 13(4), pp. 549-568.
doi: 10.1080/15583058.2018.1440030.
[11] Kron Morelli, C. (2012) I Sassi di Matera: Il percorso per la rivalutazione di una
‘Vergogna Nazionale’. Università Ca’Foscari Venezia. Thesis. Available at http://
dspace.unive.it/handle/10579/1980. [Accessed: 12th July, 2021].
[12] Neubauer, T. A. et al. (2015) A gastropod-based biogeographic scheme for the
European Neogene freshwater systems, Earth-Science Reviews. Elsevier B.V.,
143, pp. 98-116. doi: 10.1016/j.earscirev.2015.01.010.
[13] Bernardo, G. and Andrisani, G. (2016) Plaster and stone surface restoration of
ancient convicino (Sasso Caveoso in Matera, Italy), In Villegas, L. (ed) Rehabend
2016: Construction pathology, rehabilitation, technology and heritage manage-
ment. Burgos, Spain, March 2016, pp. 295-302.
[14] Fuentes, D. D. et al. (2019) Preliminary Seismic Damage Assessment of Mexican
Churches after September 2017 Earthquakes, International Journal of Architectural
Heritage. Taylor & Francis, 0(0), pp. 1-21. doi: 10.1080/15583058.2019. 1628323.
[15] D’Amato, M., Laterza, M. and Diaz Fuentes, D. (2020) Simplified Seismic Analyses
of Ancient Churches in Matera’s Landscape, International Journal of Architectural
Heritage. Taylor & Francis, 14(1), pp. 119-138. doi: 10.1080/15583058.2018.1511000.

275
 [16] Colonna, A. and Fiore, D. (2014) I Sassi e il Parco delle Chiese Rupestri di Ma-
tera Patrimonio Dell’Umanità. Piano di Gestione. Matera. Available at: https://
www.regione.basilicata.it/giunta/files/docs/DOCUMENT_FILE_2992928.pdf
[Accessed: 12th July, 2021].
[17] Fallacara, G., Barberio, M. and Colella, M. (2016) Mater[i]a: conoscenza e pro-
getto. 1st edn. Matera: Liantonio Editore.
[18] Fatiguso, F. et al. (2017) Resilience of Historic Built Environments: Inherent
Qualities and Potential Strategies, Procedia Engineering. 180, pp. 1024-1033.
doi: 10.1016/j.proeng.2017.04.262.
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

[19] Girard, L. F., Nocca, F. and Gravagnuolo, A. (2019) Matera: City of nature, city of
culture, city of regeneration. Towards a landscape-based and culture-based ur-
ban circular economy, Aestimum, 74(776758), pp. 5-42. doi: 10.13128/aestim-7007.
[20] Toxey,A. P. (2016) Materan Contradictions. Routledge. doi: 10.4324/9781315594279.
[21] World Heritage Convention UNESCO (2020) The Sassi and the Park of the Rup-
estrian Churches of Matera, The List of World Heritage. Available at: https://whc.
unesco.org/en/list/670/documents [Accessed: 12th July, 2021].
[22] Pontrandolfi P., Manganelli B. (2018) Urban Regeneration for a Sustainable and
Resilient City: An Experimentation in Matera. In: Gervasi O. et al. (eds) Compu-
tational Science and Its Applications – ICCSA 2018. ICCSA 2018. Lecture Notes
in Computer Science, vol 10964. Springer, Cham. https://doi.org/10.1007/978-3-
319-95174-4_3.
[23] Fontana, A. (2014) Inedita scoperta a Matera: Chiesa rupestre di San Pietro de
Morrone, Sassienatura, 1st July, 2013. Available at http://www.sassienatura.it/
san%20pietro%20de%20morrone%20matera.htm [Accessed: 12th July, 2021].
[24] Fontana, A. and Paolicelli, R. (2014) La chiesa rupestre di San Pietro de Morrone.
Scoperta e studio di un luogo di culto medioevale. 1st edn. Matera: Edizioni Gian-
natelli.
[25] Centonze, S. (2019) Il ritratto realistico nei graffiti della Cattedrale di Matera.
Mathera. Associazione Culturale ANTROS. III-8 pp.113-122
[26] L. Caro, J. (2012) Fotogrametría modelado 3D: un caso práctico para la difusión
del patrimonio y su promoción turística, In IX Congreso Tecnologías de la Infor-
mación y las Comunicaciones. TURITEC2012, p. 15. Available at: http://hdl.han-
dle.net/10630/5134 [Accessed: 12th July, 2021].
[27] Lerma, J. L. et al. (2013) Photorealistic 3D modelling by photogrammetry: laser
scanning versus digital image. Cuadernos de arte rupestre, 6(2013) Available
at: http://cuadernosdearterupestre.es [Accessed: 12th July, 2021].
[28] Habib, A. F., Ghanma, M. S. and Kim, E. M. (2004) 3D Modelling of Historical Sites
using LowCost Digital Cameras”. In XXXVth Congress of ISPRS proceedings.
[29] Historic England (2017) Photogrammetric Applications for Cultural Heritage.
1st edn, Guidance for Good Practice. 1st edn. Edited by J. Bedford. Swindon:
Historic England.

Biographical notes

Rafael Ramírez Eudave is a Ph. D. candidate at the Institute for Sustainability and
Innovation in Structural Engineering of the University of Minho, Portugal. He is a mem-
ber of the INFRARISK program, supported by the Portuguese Foundation for Science
and Technology. He obtained his degree as an Architect in 2015 (National Autonomous

276
University of Mexico), and was part of the Erasmus ELARCH program, later obtaining
a M. Sc. degree in Structural Analysis of Historical Constructions (University of Minho,
Portugal, 2017) and in Heritage and Project (University of Basilicata, Italy, 2018). He
has worked in the National Coordination of Historical Monuments of the Mexican Na-
tional Institute for Anthropology and History and participated in interventions on histor-
ical constructions damaged during the 2017 earthquakes in Mexico. His current re-
search is devoted to the assessment and mitigation of the seismic vulnerability of
historical centres. He is currently a member of the Order of Engineers (Portugal).

Michele D’Amato is an Assistant Professor in Structural Engineering at DiCEM,

Dept. of European and Mediterranean Cultures at the University of Basilicata, Matera,
Italy, where he teaches “Mechanics of Materials” and “Seismic Design of Structures”,
on the Architecture study course. He earned his PhD in “Mathematical Methods and
Models for Dynamic Systems” at the University of Basilicata in 2009. He has participat-
ed and continues to participate in different research projects focused on the seismic
protection of existing and historical constructions. He is the author/co-author of sever-
al international scientific papers and has participated as a lecturer in different national
and international conferences. He is a reviewer of many international scientific journals
which address structural engineering.

Giuseppe Andrisani is an Assistant Professor in Structural Engineering at DiCEM,

Dept. of European and Mediterranean Cultures of the University of Basilicata, Matera,
Italy. He is an Architect at the Istituto Universitario di Architettura in Venice, Italy; a
member of ICOMOS Italy (2011) and the Italian Castelli Institute. He has participated
and continues to participate in different research projects focused on cultural heritage
and historical constructions in Europe and Latin America. He has worked in several in-
ternational groups and conferences. He is currently working on research addressing
fortresses and military heritage.

CONSERVATION SCIENCE IN CULTURAL HERITAGE

Summary

Rupestrian cultural assets are remarkable examples of human occupation of the

territory. As for any other cultural asset, the characterisation of rupestrian buildings is
critical for their conservation and protection. However, there are a number of difficul-
ties and challenges that are inherent in the physical singularities of these architecton-
ic constructions, namely related to their geometrical irregularities. It is therefore im-
portant to use an appropriate approach to describe the geometry of these assets by
overcoming some limitations that conventional measurement and surveying ap-
proaches have. The production of digital three-dimensional models is consequently a
very useful and promising approach in this context. For most cultural assets and in
particular geometrically complex objects, the approach is useful in reconstructing
geometrical inner or exterior surfaces by means of photographic information. Photo-
grammetry is a technique used to acquire valuable and accurate three-dimensional
information, which generates reliable models that also include integrated works, such
as paintings or sculptures. The present work explores the suitability of this approach
in the context of a remarkable case of rupestrian architecture, that of the church of
San Pietro da Morrone. The building, located in the historical city of Matera (Italy), wit-
nessed a series of transformations and varied uses, before it was abandoned in the

277
 middle of the 20th century. The experimental survey of this building provides a first in-
sight into its geometry and a point of reference for discussing the implementation of a
photogrammetric approach, its advantages, limitations and possibilities.

Riassunto

I beni culturali rupestri sono esempi notevoli di occupazione umana del territorio.
Come per ogni altro bene culturale, la caratterizzazione degli edifici rupestri è fonda-
R. Ramírez Eudave, M. D’Amato, G. Andrisani - The rupestrian church of San Pietro da Morrone (Matera, Italy)

mentale per la loro conservazione e tutela. Tuttavia vi sono una serie di difficoltà e sfi-
de che sono inerenti alle singolarità fisiche di queste costruzioni architettoniche, in par-
ticolare legate alle loro irregolarità geometriche. È quindi importante utilizzare un
approccio specifico per descrivere la geometria di questi beni superando alcune limi-
tazioni che hanno gli approcci convenzionali di misurazione e rilevamento. La produ-
zione di modelli tridimensionali digitali è quindi un metodo molto utile e promettente in
questo contesto. Per la maggior parte dei beni culturali e in particolare per gli oggetti
geometricamente complessi, l’approccio è utile per ricostruire superfici geometriche in-
terne o esterne per mezzo di informazioni fotografiche.
La fotogrammetria è una tecnica utilizzata per acquisire informazioni tridimensio-
nali preziose e accurate, che genera modelli affidabili che includono anche opere inte-
grate, come dipinti o sculture. Il presente lavoro esplora l’adeguatezza di questo ap-
proccio nel contesto di un caso notevole di architettura rupestre, quello della chiesa di
San Pietro da Morrone. L’edificio, situato nella storica città di Matera (Italia), ha subito
una serie di trasformazioni e vari usi, prima di essere abbandonato a metà del XX se-
colo. Il rilievo sperimentale di questo edificio fornisce una prima visione della sua geo-
metria e un punto di riferimento per discutere l’implementazione di un approccio foto-
grammetrico, i suoi vantaggi, limiti e possibilità.

278